DEV Community

Cover image for CV-based self-diagnosis telemedicine application
Abto Software
Abto Software

Posted on

CV-based self-diagnosis telemedicine application

This post is a short overview of an Abto Software healthcare project.

Markerless human pose detection to benefit physical therapy – project overview

Our client is a health-focused organization that provides comprehensive products for the healthcare industry. The portfolio of the reputable company comprises solutions that benefit healthcare institutions and patients, from enterprise-level management platforms to user-friendly mobile applications.

Our partner was looking for knowledge and experience in delivering human body pose detection and analysis (our former successful cooperation was focused around performing personal medical device integration). Having expertise in building and implementing advanced solutions for accurate pose estimation and analysis, our company smoothly upgraded another solution.

Sensorless human body detection to transform exercise sessions: Our approach

Our team has entered into cooperation to modernize a custom telemedicine application supporting specialists in improving patient outcomes by enhancing personalized treatment and better-tailored exercise guidance. Our engineers quickly designed a solution that processes different exercises, for example, cervical flexion.

Abto Software’s main goal:

  • Initial investigation and identification of the best techniques for accurate motion tracking and analysis to enable MSK telerehabilitation
  • Efficient customization and implementation of the chosen technique for precise movement assessment to empower therapeutics monitoring

At the first stage, we trained the CV-algorithm to recognize human motion on ready-captured video materials. During the next stages, we implemented additional CV-algorithms to process human movement in real-time, allowing end-users to leverage helpful feedback during exercising.

Abto Software took over:

  • Business logic development
  • Demo product design
  • Computer vision technique implementation a. In-depth research b. Application prototyping
  • UI/UX design

CV self-diagnosis telemedicine app: Our solution

Our application was created to accelerate physical therapy:

  • The system guides patients through exercises automatically tracking every movement
  • The system, after analyzing patient performance and adherence, transfers processed health indicators to the attending clinicians who can then adjust the prescribed treatment program

The application is intended to streamline different domains:

  • Digital rehabilitation
  • Digital therapeutics
  • Spinal rehab
  • Sports medicine & orthopedics

CV camera-based telemedicine app: The challenges

Measurement points

At the discovery phase, we had to determine the correct measurement points for accurate limb assessment. This challenge was resolved by monitoring the ear-nose line segment.

Viewpoint variation

We faced viewpoint variation, which means that recognized shapes change and alter determined features. That’s why our engineers have used different datasets to train the algorithms.

Pose variation

Another issue – pose variation, which means the recognized objects aren’t steady bodies and can be deformed. To handle this problem, our engineers made sure the datasets included numerous possible variations of pose.

Summing up

Abto Software has joined the project to assist our client, a mature healthcare-focused organization in providing a user-friendly telemedicine application to empower physical therapists and patients undergoing rehabilitation. Our experts have utilized advanced technology, in particular computer vision, to facilitate remote monitoring and transform healthcare delivery.

By leveraging:

  • Artificial intelligence, ML, DL, data analytics
  • Computer vision

We assist:

  • Healthcare businesses that prioritize patient-first care
  • Physical therapists embracing innovation

And benefit:

  • Physiotherapy patients undergoing physiotherapy and rehabilitation
  • Those patients who suffer chronic conditions and prefer accessible services

Top comments (0)